Essays on Public Policy in the Informal Sector Context

206  Download (0)

Full text

(1)

ECONOMIC STUDIES

DEPARTMENT OF ECONOMICS

SCHOOL OF BUSINESS, ECONOMICS AND LAW

UNIVERSITY OF GOTHENBURG

248

________________________

Essays on Public Policy in the

Informal Sector Context

(2)

ISBN 978-91-88199-55-3 (printed) ISBN 978-91-88199-56-0 (pdf) ISSN 1651-4289 (printed) ISSN 1651-4297 (online) Printed in Sweden, Gothenburg University 2021

Till Henry och Ingrid

(3)

ISBN 978-91-88199-55-3 (printed) ISBN 978-91-88199-56-0 (pdf) ISSN 1651-4289 (printed) ISSN 1651-4297 (online)

Printed in Sweden,

(4)

1

Acknowledgments

This thesis would not have been possible without the many wonderful people who have helped me throughout this process.

First, I want to thank my supervisors M˚ans S¨oderbom and Annika Lindskog. Your knowledge, encouragement and insightful comments have guided me through the long process of writing this thesis. M˚ans, your expertise and intuitive feeling for economics and econometrics has been a true inspiration. Thank you for being my mentor and for challenging me to become a better researcher. Annika, you have an eye for detail that makes you a great researcher. Thank you for always asking the hard questions and for the countless hours your have dedicated to my work. Your efforts have challenged me to push this work further. Thank you both for everything you have taught me. It has been a pleasure working with you both. Finally, I would be remiss if I did not also thank you for being understanding. Life has its ups and downs, regardless of the demands of a PhD. During this process, I have enjoyed both the best and the worst moments in my personal life. Without your patience, guidance and support I would not have been able to finish.

My PhD journey would not have been the same without all my wonderful colleagues and cohort friends. Thank you Josephine, Lisa, Hanna, Martin, Laura, Mikael, Tensay, Vivi, Verena, Simon, Andrea, and Yashoda, for all the late night dinners in L1 (that seems like a lifetime ago), and all the support and good times we shared. I would not have made it through those first years without all of you. I would also like to thank the ”younger” phd cohorts for including me in their community when I returned from parental leave. Thank you Melissa, Simon, Tewodros, Eyoual, Sebastian, Louise, Tam´s, Maks, and Lina for always being supportive.

Lisa, thank you for all the morning and afternoon coffee breaks, and our discussions about life and econometrics. Our friendship began in L1, but it will be a life-long journey. Josephine, you have become like a sister to me. Thank you for teaching me about policy, for laughing and crying with me, and for spoiling Henry and Ingrid with the love of an aunt. You have a special place in our hearts. Anja, you have encouraged me to believe in myself, to stop whining, and pushed me to continually try harder. Thank you for the comradery of shared research and conference trips, and for inviting me to Barnard. Most of all, thank you for being my friend.

I am greatly indebted to all my colleagues at the Department of Economics for providing a great academic and collegial environment. It has been a pleasure to work

with each of you. Thank you Joe Vecci for being a good friend and for always taking the time to listen to my concerns. I truly appreciated your willingness to help me to solve whatever obstacles I confronted. A heartfelt thank you to Eva Ranehill for sharing her job market experiences with me, but more so for the understanding and offerings of help during moments when combining academic work and motherhood was challenging. It meant so much. A special thank you to Anna Bindler and Gustav Kjellson for their insightful comments during my final seminars, I am very grateful. Gustav, I especially enjoyed our interesting discussions on adverse selection and health economics. Thank you also to the development group for all the invaluable seminar discussions. I am indebted to Elizabeth F¨oldi, Mona J¨onefors, Ann-Christin R¨a¨at¨ari Nystr¨om, and Maria Siirak, for all the administrative support they provided in helping me sort out my endless questions related to parental leave. Emma, thank you for supporting my teaching assistant duties while I was in the US and for all the conversations in your office. Selma, thank you for always keeping me updated on the latest Netflix series, and Carmen, thank you for letting us stay in your home.

My research on Rwandan health policy would not have been possible without the help and support from Andinet Woldemichael at the African Development Bank. I am grateful for all your good advice and insightful guidance related to public policy and econometrics. Thank you for being a mentor and a friend. It has been my pleasure to work with you, I hope we will have the opportunity to meet soon.

Empirical research would be impossible without the knowledge and expertise of local institutions and policy makers. I am indebted to everyone at the Ministry of Health in Rwanda. Andrew Muhire, Dr. Pascal Kayobotsi, and Emmanuel Ntawuyirusha, thank you for your wonderful reception at the Ministry of Health in Kigali and for sharing you expert knowledge on the HMIS and PBF data. I am humbled by your generosity and your tireless dedication to developing the Rwandan health sector.

The Department of Social Sciences and the Collage of Business at Michigan Tech-nological University have been my home away from home during my PhD. To all of you at Michigan Tech, thank you for the welcoming atmosphere, great research con-versations and for all the cross-country skiing. Michigan is my second home and you have helped me make each trip there a pleasure.

(5)

1

Acknowledgments

This thesis would not have been possible without the many wonderful people who have helped me throughout this process.

First, I want to thank my supervisors M˚ans S¨oderbom and Annika Lindskog. Your knowledge, encouragement and insightful comments have guided me through the long process of writing this thesis. M˚ans, your expertise and intuitive feeling for economics and econometrics has been a true inspiration. Thank you for being my mentor and for challenging me to become a better researcher. Annika, you have an eye for detail that makes you a great researcher. Thank you for always asking the hard questions and for the countless hours your have dedicated to my work. Your efforts have challenged me to push this work further. Thank you both for everything you have taught me. It has been a pleasure working with you both. Finally, I would be remiss if I did not also thank you for being understanding. Life has its ups and downs, regardless of the demands of a PhD. During this process, I have enjoyed both the best and the worst moments in my personal life. Without your patience, guidance and support I would not have been able to finish.

My PhD journey would not have been the same without all my wonderful colleagues and cohort friends. Thank you Josephine, Lisa, Hanna, Martin, Laura, Mikael, Tensay, Vivi, Verena, Simon, Andrea, and Yashoda, for all the late night dinners in L1 (that seems like a lifetime ago), and all the support and good times we shared. I would not have made it through those first years without all of you. I would also like to thank the ”younger” phd cohorts for including me in their community when I returned from parental leave. Thank you Melissa, Simon, Tewodros, Eyoual, Sebastian, Louise, Tam´s, Maks, and Lina for always being supportive.

Lisa, thank you for all the morning and afternoon coffee breaks, and our discussions about life and econometrics. Our friendship began in L1, but it will be a life-long journey. Josephine, you have become like a sister to me. Thank you for teaching me about policy, for laughing and crying with me, and for spoiling Henry and Ingrid with the love of an aunt. You have a special place in our hearts. Anja, you have encouraged me to believe in myself, to stop whining, and pushed me to continually try harder. Thank you for the comradery of shared research and conference trips, and for inviting me to Barnard. Most of all, thank you for being my friend.

I am greatly indebted to all my colleagues at the Department of Economics for providing a great academic and collegial environment. It has been a pleasure to work

with each of you. Thank you Joe Vecci for being a good friend and for always taking the time to listen to my concerns. I truly appreciated your willingness to help me to solve whatever obstacles I confronted. A heartfelt thank you to Eva Ranehill for sharing her job market experiences with me, but more so for the understanding and offerings of help during moments when combining academic work and motherhood was challenging. It meant so much. A special thank you to Anna Bindler and Gustav Kjellson for their insightful comments during my final seminars, I am very grateful. Gustav, I especially enjoyed our interesting discussions on adverse selection and health economics. Thank you also to the development group for all the invaluable seminar discussions. I am indebted to Elizabeth F¨oldi, Mona J¨onefors, Ann-Christin R¨a¨at¨ari Nystr¨om, and Maria Siirak, for all the administrative support they provided in helping me sort out my endless questions related to parental leave. Emma, thank you for supporting my teaching assistant duties while I was in the US and for all the conversations in your office. Selma, thank you for always keeping me updated on the latest Netflix series, and Carmen, thank you for letting us stay in your home.

My research on Rwandan health policy would not have been possible without the help and support from Andinet Woldemichael at the African Development Bank. I am grateful for all your good advice and insightful guidance related to public policy and econometrics. Thank you for being a mentor and a friend. It has been my pleasure to work with you, I hope we will have the opportunity to meet soon.

Empirical research would be impossible without the knowledge and expertise of local institutions and policy makers. I am indebted to everyone at the Ministry of Health in Rwanda. Andrew Muhire, Dr. Pascal Kayobotsi, and Emmanuel Ntawuyirusha, thank you for your wonderful reception at the Ministry of Health in Kigali and for sharing you expert knowledge on the HMIS and PBF data. I am humbled by your generosity and your tireless dedication to developing the Rwandan health sector.

The Department of Social Sciences and the Collage of Business at Michigan Tech-nological University have been my home away from home during my PhD. To all of you at Michigan Tech, thank you for the welcoming atmosphere, great research con-versations and for all the cross-country skiing. Michigan is my second home and you have helped me make each trip there a pleasure.

(6)

you can know. Amanda, I am grateful for your unconditional friendship, for always looking after me and for offering your help when I needed it the most. We have come a long way since our days at Handels working on our master’s thesis. You inspire me every day.

Above all, I owe my deepest debt of gratitude to my family. To my mother, for her endless love and encouragement. Sebastian and Rebecca, Ann-Kristin, Johanna, and now Anna, your love, patience and unconditional support have meant everything to me. Thank you all for helping me to achieve my goal and for always having my back. Julie, Amy, Nash, and Pax, you have welcomed me into your family with open arms and I thank you. Dad, you inspired me to start this PhD, to always do my best and to be proud of myself. I know that you would have been so proud of this thesis. I miss you every day.

Finally, I would like to thank my wonderful husband Sam. You have been my biggest supporter during this journey. Thank you for helping to carry my load, for always listening, encouraging me, and challenging me to see the bigger picture. You are the love of my life and I owe you everything.

Henry and Ingrid, I love you more than words can express. This thesis is for you.

(7)

you can know. Amanda, I am grateful for your unconditional friendship, for always looking after me and for offering your help when I needed it the most. We have come a long way since our days at Handels working on our master’s thesis. You inspire me every day.

Above all, I owe my deepest debt of gratitude to my family. To my mother, for her endless love and encouragement. Sebastian and Rebecca, Ann-Kristin, Johanna, and now Anna, your love, patience and unconditional support have meant everything to me. Thank you all for helping me to achieve my goal and for always having my back. Julie, Amy, Nash, and Pax, you have welcomed me into your family with open arms and I thank you. Dad, you inspired me to start this PhD, to always do my best and to be proud of myself. I know that you would have been so proud of this thesis. I miss you every day.

Finally, I would like to thank my wonderful husband Sam. You have been my biggest supporter during this journey. Thank you for helping to carry my load, for always listening, encouraging me, and challenging me to see the bigger picture. You are the love of my life and I owe you everything.

Henry and Ingrid, I love you more than words can express. This thesis is for you.

Gothenburg, May 2021 Carolin Sj¨oholm

Contents

Acknowledgments

i

Introduction

1

I: The Price Sensitivity of Demand for Health Insurance: Evidence

from Community Based Health Insurance in Rwanda

1

Introduction . . . .

1

2

Community-Based Health Insurance . . . .

6

3

Data . . . 11

4

Empirical Strategy . . . 13

5

Results . . . 21

6

Financial Self-sustainability . . . 35

7

Conclusions . . . 44

Appendices . . . 54

II: The Role of Childcare in Firm Performance: Evidence from

Female Entrepreneurship in Mexico

1

Introduction . . . .

1

2

Childcare in Mexico . . . .

5

3

Data . . . .

7

4

Empirical Strategy . . . 14

5

Results . . . 20

6

Conclusion

. . . 35

Appendices . . . 44

III:Variation in the Quality of Primary Healthcare: Evidence from

Rural and Urban Health Services in Rwanda

1

Introduction . . . .

1

2

The Rwandan Health Sector . . . .

4

3

Data . . . .

7

4

Empirical Analysis . . . 15

5

Results . . . 16

6

Conclusion

. . . 24

(8)

Introduction

A majority of all workers in the world are informally employed. Approximately 2 billion workers, or 60% of the world’s employed population, earn their livelihoods in the infor-mal sector. Workers in the inforinfor-mal sector often face higher risk of poverty and lower productivity compared to formal workers (International Labour Office, 2018). One reason is that they often lack access to social protection, which makes them vulnerable to adverse shocks such as sickness or income loss. Social protection refers to policies and programs designed to reduce and prevent poverty and vulnerability throughout the life cycle and often include a mix of social insurance and social assistance programs (International Labour Office, 2017). In the absence of these safety nets, adverse shocks risk pushing households deeper into poverty or maintain them in a poverty trap.

During the last decades, large efforts have been made in many developing countries to expand social protection to the informal sector and achieve universal coverage of such programs. Despite wide agreement regarding the importance of social protection as a key factor for inclusive growth, this human right is still not fulfilled for most people in the world. The 1948 Universal Declaration of Human Rights recognizes that everyone has the right to social security (Article 22). Although social protection policies are seen as key elements in national development strategies in most countries, it has been estimated that approximately 4 billion people, almost 55% of the world’s population, have access to no or inadequate social protection (International Labour Office, 2017). A majority of this group is represented by households in the informal sector.

Households in the informal sector face substantial idiosyncratic and common risk, resulting in high income variability (Townsend, 1994). For a large share of these house-holds who live on a day to day basis, adverse shocks such as health and employment shocks, could throw families into poverty and have long lasting effects for generations. For example, in 2015 approximately 930 million people incurred catastrophic health expenditures, defined as out-of-pocket health spending exceeding 10% of household consumption, pushing 89.7 million people into extreme poverty (World Health Organi-zation & World Bank, 2019). As a result, informal labor is increasingly recognized as an obstacle to eradicate poverty in many developing countries and a major challenge for achieving the Sustainable Development Goals.

(9)

Introduction

A majority of all workers in the world are informally employed. Approximately 2 billion workers, or 60% of the world’s employed population, earn their livelihoods in the infor-mal sector. Workers in the inforinfor-mal sector often face higher risk of poverty and lower productivity compared to formal workers (International Labour Office, 2018). One reason is that they often lack access to social protection, which makes them vulnerable to adverse shocks such as sickness or income loss. Social protection refers to policies and programs designed to reduce and prevent poverty and vulnerability throughout the life cycle and often include a mix of social insurance and social assistance programs (International Labour Office, 2017). In the absence of these safety nets, adverse shocks risk pushing households deeper into poverty or maintain them in a poverty trap.

During the last decades, large efforts have been made in many developing countries to expand social protection to the informal sector and achieve universal coverage of such programs. Despite wide agreement regarding the importance of social protection as a key factor for inclusive growth, this human right is still not fulfilled for most people in the world. The 1948 Universal Declaration of Human Rights recognizes that everyone has the right to social security (Article 22). Although social protection policies are seen as key elements in national development strategies in most countries, it has been estimated that approximately 4 billion people, almost 55% of the world’s population, have access to no or inadequate social protection (International Labour Office, 2017). A majority of this group is represented by households in the informal sector.

Households in the informal sector face substantial idiosyncratic and common risk, resulting in high income variability (Townsend, 1994). For a large share of these house-holds who live on a day to day basis, adverse shocks such as health and employment shocks, could throw families into poverty and have long lasting effects for generations. For example, in 2015 approximately 930 million people incurred catastrophic health expenditures, defined as out-of-pocket health spending exceeding 10% of household consumption, pushing 89.7 million people into extreme poverty (World Health Organi-zation & World Bank, 2019). As a result, informal labor is increasingly recognized as an obstacle to eradicate poverty in many developing countries and a major challenge for achieving the Sustainable Development Goals.

(10)

borrowing, asset sale and decreased education expenditures (Leive & Xu, 2008; Mitra et al., 2016; Heltberg & Lund, 2009; Islam & Maitra, 2012). Additionally, uninsured risk compels households to diversify income and to engage in low-risk and low-return production activities (Cole et al., 2017; Dercon & Christiaensen, 2011) in order to smooth consumption. These activities hamper the ability of households to grow their incomes and escape poverty (Binswanger & Rosenzweig, 1993). As a result, households are kept in persistent poverty. Despite informal insurance arrangements and strate-gies income fluctuations often remain high, suggesting that informal income-smoothing mechanisms are inadequate and leave households with uninsured risk (Townsend, 1995; P. Gertler & Gruber, 2002).

Social protection can reduce the cost of coping strategies and enhance the capacity of families and communities to absorb the negative impacts of shocks. For example, cash transfers have been shown to have positive and sustained effects on child education and health (Aizer et al., 2016; Baird et al., 2019), household investment in durable goods and savings (Haushofer & Shapiro, 2016), and productive investments (Handa et al., 2018; Bastagli et al., 2016; P. J. Gertler et al., 2012). Furthermore, in the presence of idiosyncratic shocks, households with health insurance were more likely to invest in schooling for girls, livestock and durable goods compared to uninsured households (Liu, 2016). There is an international consensus on the importance of social protection as a key policy tool for implementing the U.N. Sustainable Development Goals and to ensure inclusive development where no one is left behind. Social protection is essential to achieving a number of the SGDs such as eradicating poverty for all everywhere (SDG 1), ending hunger (SDG 2), and contributing to gender equity and women’s empowerment (SDG 5). Furthermore, by increasing access to affordable healthcare, social protection can contribute to achieving universal health care (Target 3.8) and good health and well-being for all (SDG 3). As a result, Target 1.3 explicitly calls on countries to implement nationally appropriate social protection systems to end poverty by 2030 (United Nations, 2015).

Despite a global agreement on the importance of social protection, questions regard-ing how to best implement and expand effective and sustainable universal programs are still unanswered. Countries often combine contributory social insurance schemes with non-contributory social assistance programs in order to achieve a universal coverage. On the one hand, non-contributory programs include universal and means-tested social assistance programs that are key to ensuring a basic level of social protection for all

residents, i.e. a social protection floor (Behrendt & Nguyen, 2018). While universal programs are effective in reaching the poorest and most vulnerable households, they also cover many households that are not in need of social protection. In the face of lim-ited fiscal capacities, programs targeted to the poor might offer a more cost-effective option. However, means-tested programs rely on costly mechanisms to identify the poorest households (Aryeetey et al., 2012) with low levels of accuracy. This often leads to under-coverage and errors of exclusion (Brown et al., 2016), resulting in a trade-off between coverage and effectiveness.

On the other hand, contributory social insurance schemes tend to provide more insurance coverage and a higher level of protection than social assistance programs. However, social insurance schemes might be inaccessible for the poorest households that often lack contributory capacity (Behrendt & Nguyen, 2018). In order to make enrollment equitable, governments can subsidize enrollment. However, subsidies have the potential of being regressive if contributions remain too high for the most vulnerable households, preventing them from enrollment despite government subsidies (Kalisa et al., 2016). Additionally, take-up of the social protection programs might be hindered by factors such as lack of information (Hossain, 2011), high transaction costs (Capuno et al., 2016), and low quality of services. These barriers must be defined and targeted by well designed policies and interventions.

Ultimately, the potential capacity of social protection programs to address risk and vulnerability, by contributing to increased productivity and resilience among house-holds in the informal sector, represents another important factor that is likely to predict take-up of the program and willingness to contribute to enrollment. This is largely con-textual. If benefits are not aligned with the need and priorities of households, they may be reluctant to contribute. The design of efficient policies is likely to be particularly challenging for the informal sector that represents a complex and all but homogeneous sector of the labor force in most developing country contexts. Increased knowledge regarding the impacts of social protection programs can improve the predictability and the efficiency of public policy.

Summary Thesis

(11)

borrowing, asset sale and decreased education expenditures (Leive & Xu, 2008; Mitra et al., 2016; Heltberg & Lund, 2009; Islam & Maitra, 2012). Additionally, uninsured risk compels households to diversify income and to engage in low-risk and low-return production activities (Cole et al., 2017; Dercon & Christiaensen, 2011) in order to smooth consumption. These activities hamper the ability of households to grow their incomes and escape poverty (Binswanger & Rosenzweig, 1993). As a result, households are kept in persistent poverty. Despite informal insurance arrangements and strate-gies income fluctuations often remain high, suggesting that informal income-smoothing mechanisms are inadequate and leave households with uninsured risk (Townsend, 1995; P. Gertler & Gruber, 2002).

Social protection can reduce the cost of coping strategies and enhance the capacity of families and communities to absorb the negative impacts of shocks. For example, cash transfers have been shown to have positive and sustained effects on child education and health (Aizer et al., 2016; Baird et al., 2019), household investment in durable goods and savings (Haushofer & Shapiro, 2016), and productive investments (Handa et al., 2018; Bastagli et al., 2016; P. J. Gertler et al., 2012). Furthermore, in the presence of idiosyncratic shocks, households with health insurance were more likely to invest in schooling for girls, livestock and durable goods compared to uninsured households (Liu, 2016). There is an international consensus on the importance of social protection as a key policy tool for implementing the U.N. Sustainable Development Goals and to ensure inclusive development where no one is left behind. Social protection is essential to achieving a number of the SGDs such as eradicating poverty for all everywhere (SDG 1), ending hunger (SDG 2), and contributing to gender equity and women’s empowerment (SDG 5). Furthermore, by increasing access to affordable healthcare, social protection can contribute to achieving universal health care (Target 3.8) and good health and well-being for all (SDG 3). As a result, Target 1.3 explicitly calls on countries to implement nationally appropriate social protection systems to end poverty by 2030 (United Nations, 2015).

Despite a global agreement on the importance of social protection, questions regard-ing how to best implement and expand effective and sustainable universal programs are still unanswered. Countries often combine contributory social insurance schemes with non-contributory social assistance programs in order to achieve a universal coverage. On the one hand, non-contributory programs include universal and means-tested social assistance programs that are key to ensuring a basic level of social protection for all

residents, i.e. a social protection floor (Behrendt & Nguyen, 2018). While universal programs are effective in reaching the poorest and most vulnerable households, they also cover many households that are not in need of social protection. In the face of lim-ited fiscal capacities, programs targeted to the poor might offer a more cost-effective option. However, means-tested programs rely on costly mechanisms to identify the poorest households (Aryeetey et al., 2012) with low levels of accuracy. This often leads to under-coverage and errors of exclusion (Brown et al., 2016), resulting in a trade-off between coverage and effectiveness.

On the other hand, contributory social insurance schemes tend to provide more insurance coverage and a higher level of protection than social assistance programs. However, social insurance schemes might be inaccessible for the poorest households that often lack contributory capacity (Behrendt & Nguyen, 2018). In order to make enrollment equitable, governments can subsidize enrollment. However, subsidies have the potential of being regressive if contributions remain too high for the most vulnerable households, preventing them from enrollment despite government subsidies (Kalisa et al., 2016). Additionally, take-up of the social protection programs might be hindered by factors such as lack of information (Hossain, 2011), high transaction costs (Capuno et al., 2016), and low quality of services. These barriers must be defined and targeted by well designed policies and interventions.

Ultimately, the potential capacity of social protection programs to address risk and vulnerability, by contributing to increased productivity and resilience among house-holds in the informal sector, represents another important factor that is likely to predict take-up of the program and willingness to contribute to enrollment. This is largely con-textual. If benefits are not aligned with the need and priorities of households, they may be reluctant to contribute. The design of efficient policies is likely to be particularly challenging for the informal sector that represents a complex and all but homogeneous sector of the labor force in most developing country contexts. Increased knowledge regarding the impacts of social protection programs can improve the predictability and the efficiency of public policy.

Summary Thesis

(12)

social protection programs and to increase resilience among households in the informal sector. To do this, I use quasi-experimental evaluation methods in combination with detailed administrative data to investigate the impact of policies on the take-up and quality of social insurance programs in the informal sector, as well as their impact on the productivity of economic activities. The overall contribution of my thesis is to provide evidence related to the interaction between public policy and household decision making, focusing primarily on economic performance and health. On the one hand, the design and implementation of policy interventions will determine the effects of such policies on the lives and livelihoods of households in the informal sector. On the other hand, household preferences and decision making will determine the success of implementation and take-up of these policies. In the first chapter I evaluate the importance of premium subsidies as a policy tool to achieve universal coverage of community-based health insurance (CBHI) in Rwanda, while considering the financial self-sustainability of the insurance scheme. In the second paper I explore the impact of a national social protection program in Mexico that addresses the burden of unpaid housework on women, when evaluating its impact on female entrepreneurship.Finally, in the third paper I investigate disparities in the quality of health services provided within the community-based health insurance scheme in Rwanda.

In chapter 1, ”The Price Sensitivity of Demand for Health Insurance: Evidence from Community Based Health Insurance in Rwanda”, I use the introduction of a new premium subsidy scheme to estimate the price sensitivity of the demand for community-based health insurance in Rwanda. The results indicate that the demand for health insurance is price sensitive but not elastic, suggesting that the demand for health insur-ance varies little in relation to the variation in price. Furthermore, my findings suggest that the price sensitivity varies among socioeconomic groups. I use the estimated price sensitivity to predict insurance coverage and the financial self-sustainability in relation to a number of plausible subsidy schemes. As a direct result of the inelastic insurance demand, take-up does not vary much across premium schemes. However, the heterogeneity in price sensitivity indicates that premium subsidies will affect the composition of enrolled individuals. To estimate the financial self-sustainability of the subsidy schemes, I match administrative data on the cost of providing the insurance scheme to estimate how much of the total insurer costs are covered by premiums for each alternative subsidy scheme. This allows me to control for the potential effects of adverse selection on patient costs. I find a positive slope of the cost curve, which is

consistent with adverse selection, although the estimations suggest that the financial implications of this are limited. Overall, the results suggest that premium subsidies might represent an expensive policy tool for reaching universal heath coverage, one of the key targets in the Sustainable Development Goals.

In chapter 2, ”The Role of Childcare in Firm Performance: Evidence from Female Entrepreneurship in Mexico”, I study the impact of a national daycare program on the performance of female-run microenterprises. Estancias Infantiles para Apoyar a

Madres Trabajadoras offered subsidized childcare to children younger than four years

old, whose mother was working in the informal sector. I explore the variation in availability of the program in a difference-in-difference design, and compare outcomes for women with children just below and above the eligibility threshold for the program. Furthermore, I explore the roll-out of the childcare program in a triple-difference design with treatment intensity that varies across municipalities and over time. I find no evidence that the program was associated with changes in female entrepreneurship and business performance, proxied by number of workhours, physical capital, or the likelihood of operating the business from home, having an employee, or applying for a credit. This paper is one of the first papers to study the importance of childcare obligations as a barrier for business performance among female-run microenterprises by evaluating the impact of a nationwide social policy program on female business performance.

(13)

social protection programs and to increase resilience among households in the informal sector. To do this, I use quasi-experimental evaluation methods in combination with detailed administrative data to investigate the impact of policies on the take-up and quality of social insurance programs in the informal sector, as well as their impact on the productivity of economic activities. The overall contribution of my thesis is to provide evidence related to the interaction between public policy and household decision making, focusing primarily on economic performance and health. On the one hand, the design and implementation of policy interventions will determine the effects of such policies on the lives and livelihoods of households in the informal sector. On the other hand, household preferences and decision making will determine the success of implementation and take-up of these policies. In the first chapter I evaluate the importance of premium subsidies as a policy tool to achieve universal coverage of community-based health insurance (CBHI) in Rwanda, while considering the financial self-sustainability of the insurance scheme. In the second paper I explore the impact of a national social protection program in Mexico that addresses the burden of unpaid housework on women, when evaluating its impact on female entrepreneurship.Finally, in the third paper I investigate disparities in the quality of health services provided within the community-based health insurance scheme in Rwanda.

In chapter 1, ”The Price Sensitivity of Demand for Health Insurance: Evidence from Community Based Health Insurance in Rwanda”, I use the introduction of a new premium subsidy scheme to estimate the price sensitivity of the demand for community-based health insurance in Rwanda. The results indicate that the demand for health insurance is price sensitive but not elastic, suggesting that the demand for health insur-ance varies little in relation to the variation in price. Furthermore, my findings suggest that the price sensitivity varies among socioeconomic groups. I use the estimated price sensitivity to predict insurance coverage and the financial self-sustainability in relation to a number of plausible subsidy schemes. As a direct result of the inelastic insurance demand, take-up does not vary much across premium schemes. However, the heterogeneity in price sensitivity indicates that premium subsidies will affect the composition of enrolled individuals. To estimate the financial self-sustainability of the subsidy schemes, I match administrative data on the cost of providing the insurance scheme to estimate how much of the total insurer costs are covered by premiums for each alternative subsidy scheme. This allows me to control for the potential effects of adverse selection on patient costs. I find a positive slope of the cost curve, which is

consistent with adverse selection, although the estimations suggest that the financial implications of this are limited. Overall, the results suggest that premium subsidies might represent an expensive policy tool for reaching universal heath coverage, one of the key targets in the Sustainable Development Goals.

In chapter 2, ”The Role of Childcare in Firm Performance: Evidence from Female Entrepreneurship in Mexico”, I study the impact of a national daycare program on the performance of female-run microenterprises. Estancias Infantiles para Apoyar a

Madres Trabajadoras offered subsidized childcare to children younger than four years

old, whose mother was working in the informal sector. I explore the variation in availability of the program in a difference-in-difference design, and compare outcomes for women with children just below and above the eligibility threshold for the program. Furthermore, I explore the roll-out of the childcare program in a triple-difference design with treatment intensity that varies across municipalities and over time. I find no evidence that the program was associated with changes in female entrepreneurship and business performance, proxied by number of workhours, physical capital, or the likelihood of operating the business from home, having an employee, or applying for a credit. This paper is one of the first papers to study the importance of childcare obligations as a barrier for business performance among female-run microenterprises by evaluating the impact of a nationwide social policy program on female business performance.

(14)

contributes to the discussion on how to erase inequality in health services within de-veloping countries by providing evidence suggesting that variation in structural inputs is unlikely to erase such disparities.

References

Aizer, A., Eli, S., Ferrie, J., & Lleras-Muney, A. (2016). The long-run impact of cash transfers to poor families. American Economic Review , 106 (4), 935–71.

Aryeetey, G. C., Jehu-Appiah, C., Spaan, E., Agyepong, I., & Baltussen, R. (2012). Costs, equity, efficiency and feasibility of identifying the poor in ghana’s national health insurance scheme: empirical analysis of various strategies. Tropical Medicine

& International Health, 17 (1), 43–51.

Baird, S., McIntosh, C., & ¨Ozler, B. (2019). When the money runs out: Do cash trans-fers have sustained effects on human capital accumulation? Journal of Development

Economics, 140 , 169–85.

Bastagli, F., Hagen-Zanker, J., Harman, L., Barca, V., Sturge, G., Schmidt, T., & Pellerano, L. (2016). Cash transfers: What does the evidence say? A rigorous

review of impacts and the role of design and implementation features (Tech. Rep.).

London: Overseas Development Institute.

Behrendt, C., & Nguyen, Q. A. (2018). Innovative approaches for ensuring universal

social protection for the future of work (Future of Work Research Paper Series No. 1).

Geneva: International Labour Organization.

Binswanger, H., & Rosenzweig, M. (1993). Wealth, weather risk and the composition and profitability of agricultural investments. Economic Journal , 103 (416), 56–78. Brown, C., Ravallion, M., & Van de Walle, D. (2016). A poor means test? econometric

targeting in africa (Working Paper No. 22919). Cambridge, MA, USA: National

Bureau of Economic Research.

Capuno, J. J., Kraft, A. D., Quimbo, S., Tan Jr, C. R., & Wagstaff, A. (2016). Effects of price, information, and transactions cost interventions to raise voluntary enrollment in a social health insurance scheme: A randomized experiment in the Philippines.

Health economics, 25 (6), 650–62.

Cole, S., Gin´e, X., & Vickery, J. (2017). How does risk management influence produc-tion decisions? Evidence from a field experiment. The Review of Financial Studies,

30 (6), 1935–70.

Dercon, S., & Christiaensen, L. (2011). Consumption risk, technology adoption and poverty traps: Evidence from ethiopia. Journal of development economics, 96 (2), 159–73.

Gertler, P., & Gruber, J. (2002). Insuring consumption against illness. American

economic review , 92 (1), 51–70.

Gertler, P. J., Martinez, S. W., & Rubio-Codina, M. (2012). Investing cash transfers to raise long-term living standards. American Economic Journal: Applied Economics,

4 (1), 164–92.

Handa, S., Natali, L., Seidenfeld, D., Tembo, G., Davis, B., & Team, Z. C. T. E. S. (2018). Can unconditional cash transfers raise long-term living standards? Evidence from Zambia. Journal of Development Economics, 133 , 42–65.

Haushofer, J., & Shapiro, J. (2016). The short-term impact of unconditional cash transfers to the poor: experimental evidence from kenya. The Quarterly Journal of

Economics, 131 (4), 1973–2042.

Heltberg, R., & Lund, N. (2009). Shocks, coping, and outcomes for pakistan’s poor: health risks predominate. The Journal of Development Studies, 45 (6), 889–910. Hossain, Z. (2011). Extreme poor adivasis and the problem of accessing social safety

nets (Shiree Working Paper No. 4). Dhaka, Bangladesh: Shiree, Extreme Poverty

Research Group.

International Labour Office. (2017). World Social Protection Report 2017–2019:

Uni-versal social protection to achieve the Sustainable Development Goals (World Social

Protection Report). Geneva: International Labour Organization.

International Labour Office. (2018). Women and men in the informal economy: A

statistical picture (Tech. Rep.). Geneva: International Labour Organization.

(15)

contributes to the discussion on how to erase inequality in health services within de-veloping countries by providing evidence suggesting that variation in structural inputs is unlikely to erase such disparities.

References

Aizer, A., Eli, S., Ferrie, J., & Lleras-Muney, A. (2016). The long-run impact of cash transfers to poor families. American Economic Review , 106 (4), 935–71.

Aryeetey, G. C., Jehu-Appiah, C., Spaan, E., Agyepong, I., & Baltussen, R. (2012). Costs, equity, efficiency and feasibility of identifying the poor in ghana’s national health insurance scheme: empirical analysis of various strategies. Tropical Medicine

& International Health, 17 (1), 43–51.

Baird, S., McIntosh, C., & ¨Ozler, B. (2019). When the money runs out: Do cash trans-fers have sustained effects on human capital accumulation? Journal of Development

Economics, 140 , 169–85.

Bastagli, F., Hagen-Zanker, J., Harman, L., Barca, V., Sturge, G., Schmidt, T., & Pellerano, L. (2016). Cash transfers: What does the evidence say? A rigorous

review of impacts and the role of design and implementation features (Tech. Rep.).

London: Overseas Development Institute.

Behrendt, C., & Nguyen, Q. A. (2018). Innovative approaches for ensuring universal

social protection for the future of work (Future of Work Research Paper Series No. 1).

Geneva: International Labour Organization.

Binswanger, H., & Rosenzweig, M. (1993). Wealth, weather risk and the composition and profitability of agricultural investments. Economic Journal , 103 (416), 56–78. Brown, C., Ravallion, M., & Van de Walle, D. (2016). A poor means test? econometric

targeting in africa (Working Paper No. 22919). Cambridge, MA, USA: National

Bureau of Economic Research.

Capuno, J. J., Kraft, A. D., Quimbo, S., Tan Jr, C. R., & Wagstaff, A. (2016). Effects of price, information, and transactions cost interventions to raise voluntary enrollment in a social health insurance scheme: A randomized experiment in the Philippines.

Health economics, 25 (6), 650–62.

Cole, S., Gin´e, X., & Vickery, J. (2017). How does risk management influence produc-tion decisions? Evidence from a field experiment. The Review of Financial Studies,

30 (6), 1935–70.

Dercon, S., & Christiaensen, L. (2011). Consumption risk, technology adoption and poverty traps: Evidence from ethiopia. Journal of development economics, 96 (2), 159–73.

Gertler, P., & Gruber, J. (2002). Insuring consumption against illness. American

economic review , 92 (1), 51–70.

Gertler, P. J., Martinez, S. W., & Rubio-Codina, M. (2012). Investing cash transfers to raise long-term living standards. American Economic Journal: Applied Economics,

4 (1), 164–92.

Handa, S., Natali, L., Seidenfeld, D., Tembo, G., Davis, B., & Team, Z. C. T. E. S. (2018). Can unconditional cash transfers raise long-term living standards? Evidence from Zambia. Journal of Development Economics, 133 , 42–65.

Haushofer, J., & Shapiro, J. (2016). The short-term impact of unconditional cash transfers to the poor: experimental evidence from kenya. The Quarterly Journal of

Economics, 131 (4), 1973–2042.

Heltberg, R., & Lund, N. (2009). Shocks, coping, and outcomes for pakistan’s poor: health risks predominate. The Journal of Development Studies, 45 (6), 889–910. Hossain, Z. (2011). Extreme poor adivasis and the problem of accessing social safety

nets (Shiree Working Paper No. 4). Dhaka, Bangladesh: Shiree, Extreme Poverty

Research Group.

International Labour Office. (2017). World Social Protection Report 2017–2019:

Uni-versal social protection to achieve the Sustainable Development Goals (World Social

Protection Report). Geneva: International Labour Organization.

International Labour Office. (2018). Women and men in the informal economy: A

statistical picture (Tech. Rep.). Geneva: International Labour Organization.

(16)

Kalisa, I., Musange, S., Collins, D., Saya, U., Kunda, T., & Parfait, U. (2016). The

de-velopment of community-based health insurance in rwanda: Experiences and lessons

(Tech. Rep.). Kigali: University of Rwanda College of Medicine and Health Sciences, School of Public Health.

Leive, A., & Xu, K. (2008). Coping with out-of-pocket health payments: Empirical evidence from 15 african countries. Bulletin of the World Health Organization, 86 , 849–56C.

Liu, K. (2016). Insuring against health shocks: Health insurance and household choices.

Journal of health economics, 46 , 16–32.

Mitra, S., Palmer, M., Mont, D., & Groce, N. (2016). Can households cope with health shocks in vietnam? Health economics, 25 (7), 888–907.

Townsend, R. M. (1994). Risk and insurance in village India. Econometrica: Journal

of the Econometric Society, 539–91.

Townsend, R. M. (1995). Consumption insurance: An evaluation of risk-bearing systems in low-income economies. Journal of Economic perspectives, 9 (3), 83–102. United Nations. (2015). Transforming our world: the 2030 Agenda for

Sus-tainable Development (A/RES/70/1). New York: United Nations. Retrieved fromhttps://sustainabledevelopment.un.org/content/documents/21252030% 20Agenda%20for%20Sustainable%20Development%20web.pdf (Accessed April 23 2021)

World Health Organization, & World Bank. (2019). Global monitoring report on

(17)

Kalisa, I., Musange, S., Collins, D., Saya, U., Kunda, T., & Parfait, U. (2016). The

de-velopment of community-based health insurance in rwanda: Experiences and lessons

(Tech. Rep.). Kigali: University of Rwanda College of Medicine and Health Sciences, School of Public Health.

Leive, A., & Xu, K. (2008). Coping with out-of-pocket health payments: Empirical evidence from 15 african countries. Bulletin of the World Health Organization, 86 , 849–56C.

Liu, K. (2016). Insuring against health shocks: Health insurance and household choices.

Journal of health economics, 46 , 16–32.

Mitra, S., Palmer, M., Mont, D., & Groce, N. (2016). Can households cope with health shocks in vietnam? Health economics, 25 (7), 888–907.

Townsend, R. M. (1994). Risk and insurance in village India. Econometrica: Journal

of the Econometric Society, 539–91.

Townsend, R. M. (1995). Consumption insurance: An evaluation of risk-bearing systems in low-income economies. Journal of Economic perspectives, 9 (3), 83–102. United Nations. (2015). Transforming our world: the 2030 Agenda for

Sus-tainable Development (A/RES/70/1). New York: United Nations. Retrieved fromhttps://sustainabledevelopment.un.org/content/documents/21252030% 20Agenda%20for%20Sustainable%20Development%20web.pdf (Accessed April 23 2021)

World Health Organization, & World Bank. (2019). Global monitoring report on

financial protection in health 2019. Retrieved fromhttps://apps.who.int/iris/ bitstream/handle/10665/331748/9789240003958-eng.pdf (Accessed on May 2 2021)

(18)

The Price Sensitivity of Demand for Health

Insurance

Evidence from Community Based Health Insurance in Rwanda

Carolin Sj¨oholm

*

Abstract

This study estimates the price sensitivity of the demand for health insurance, ex-ploiting the variation in insurance premiums created by the implementation of a new premium subsidy scheme for community- based health insurance in Rwanda. I use the estimated price elasticity to predict the impact of a number of plausible premium subsidy schemes on two policy- relevant outcomes: insurance cover-age and financial sustainability. I find that the demand for health insurance is inelastic, although the price sensitivity varies among different socioeconomic groups. The results suggest that premium subsidies have only a modest effect on the take-up of insurance compared with nonsubsidized premiums, but they affect the composition of individuals enrolled in the insurance. To simulate the financial sustainability of the insurance scheme, measured as the share of total insurer costs covered by insurance premiums, I combine the price elasticity es-timates with unique data on insurer costs, enabling me to account for adverse selection. I estimate a positive slope of the average cost curve, consistent with adverse selection. These results indicate that premium subsidies might not rep-resent a financially sustainable policy tool for achieving universal healthcare. JEL classification: I13, I18, D12, H51, H55

Keywords: community-based health insurance, adverse selection, price sensi-tivity

*University of Gothenburg, carolin.sjoholm@economics.gu.se. I would like to thank my advisors

M˚ans S¨oderbom and Annika Lindskog for their invaluable input. This paper has also benefited greatly

(19)

The Price Sensitivity of Demand for Health

Insurance

Evidence from Community Based Health Insurance in Rwanda

Carolin Sj¨oholm

*

Abstract

This study estimates the price sensitivity of the demand for health insurance, ex-ploiting the variation in insurance premiums created by the implementation of a new premium subsidy scheme for community- based health insurance in Rwanda. I use the estimated price elasticity to predict the impact of a number of plausible premium subsidy schemes on two policy- relevant outcomes: insurance cover-age and financial sustainability. I find that the demand for health insurance is inelastic, although the price sensitivity varies among different socioeconomic groups. The results suggest that premium subsidies have only a modest effect on the take-up of insurance compared with nonsubsidized premiums, but they affect the composition of individuals enrolled in the insurance. To simulate the financial sustainability of the insurance scheme, measured as the share of total insurer costs covered by insurance premiums, I combine the price elasticity es-timates with unique data on insurer costs, enabling me to account for adverse selection. I estimate a positive slope of the average cost curve, consistent with adverse selection. These results indicate that premium subsidies might not rep-resent a financially sustainable policy tool for achieving universal healthcare. JEL classification: I13, I18, D12, H51, H55

Keywords: community-based health insurance, adverse selection, price sensi-tivity

*University of Gothenburg, carolin.sjoholm@economics.gu.se. I would like to thank my advisors

M˚ans S¨oderbom and Annika Lindskog for their invaluable input. This paper has also benefited greatly

(20)

1

Introduction

In 2015 the UN General Assembly included universal health coverage as part of the overall commitment to the Sustainable Development Goals (United Nations, 2015). Community-based health insurance (CBHI) has been adopted by many developing countries as a financing mechanism to reach this goal by pooling health risks and resources at the community level.1 So far enrollment in CBHI has often been low,

particularly among poor households (Gnawali et al., 2009; Yilma et al., 2015; Parmar et al., 2014). In order to increase enrollment levels, many countries have implemented premium subsidies.2 However, premium subsidies are costly and contribute to a lack of

self-financing of the insurance schemes as premium revenues cover only a small share of the patient costs. In addition to impacts on insurance enrollment, premium subsidies might affect the type of individuals who enroll, resulting in an association between premiums and the insurer costs. Previous research from developing country contexts suggests that premium subsidies could exacerbate the effects of adverse selection (Par-mar et al., 2012), which would negatively affect the financial sustainability of the CBHI scheme.

The main contribution of this paper is to study the impact of premium subsidies on policy relevant outcomes such as insurance coverage and the financial self-sustainability of the CBHI scheme. I use the introduction of a new premium scheme as a quasi experiment to estimate the price sensitivity of demand for health insurance. This is done in the context of Rwanda, a low-middle income country in Africa. Next, I use the estimated price sensitivity to predict enrollment levels, and subsequently premium revenue, for a number of plausible premium subsidy schemes. In order to evaluate the financial sustainability associated with the different subsidy schemes, I use unique data on the total insurer costs related to the CBHI to consider the potential effects of adverse selection on the cost of providing the health insurance. To the best of my knowledge, this is the first attempt to estimate the consequences of premium subsidies on the financial sustainability of health insurance in a developing country context, considering the financial impact of adverse selection.

In 2011 the government of Rwanda replaced a uniform subsidy scheme, in which all individuals paid the same premium, with a targeted premium subsidy that was directed

1India (Aggarwal, 2010), Uganda (Basaza et al., 2008), Burkina Faso (Fink et al., 2013)

2Mexico (Bosch et al., 2012), Vietnam (Wagstaff et al., 2016), and Ghana (Asuming et al., 2017),

(21)

1

Introduction

In 2015 the UN General Assembly included universal health coverage as part of the overall commitment to the Sustainable Development Goals (United Nations, 2015). Community-based health insurance (CBHI) has been adopted by many developing countries as a financing mechanism to reach this goal by pooling health risks and resources at the community level.1 So far enrollment in CBHI has often been low,

particularly among poor households (Gnawali et al., 2009; Yilma et al., 2015; Parmar et al., 2014). In order to increase enrollment levels, many countries have implemented premium subsidies.2 However, premium subsidies are costly and contribute to a lack of

self-financing of the insurance schemes as premium revenues cover only a small share of the patient costs. In addition to impacts on insurance enrollment, premium subsidies might affect the type of individuals who enroll, resulting in an association between premiums and the insurer costs. Previous research from developing country contexts suggests that premium subsidies could exacerbate the effects of adverse selection (Par-mar et al., 2012), which would negatively affect the financial sustainability of the CBHI scheme.

The main contribution of this paper is to study the impact of premium subsidies on policy relevant outcomes such as insurance coverage and the financial self-sustainability of the CBHI scheme. I use the introduction of a new premium scheme as a quasi experiment to estimate the price sensitivity of demand for health insurance. This is done in the context of Rwanda, a low-middle income country in Africa. Next, I use the estimated price sensitivity to predict enrollment levels, and subsequently premium revenue, for a number of plausible premium subsidy schemes. In order to evaluate the financial sustainability associated with the different subsidy schemes, I use unique data on the total insurer costs related to the CBHI to consider the potential effects of adverse selection on the cost of providing the health insurance. To the best of my knowledge, this is the first attempt to estimate the consequences of premium subsidies on the financial sustainability of health insurance in a developing country context, considering the financial impact of adverse selection.

In 2011 the government of Rwanda replaced a uniform subsidy scheme, in which all individuals paid the same premium, with a targeted premium subsidy that was directed

1India (Aggarwal, 2010), Uganda (Basaza et al., 2008), Burkina Faso (Fink et al., 2013)

2Mexico (Bosch et al., 2012), Vietnam (Wagstaff et al., 2016), and Ghana (Asuming et al., 2017),

(22)

to households with low socioeconomic status. The aim of the targeted subsidy was to increase access to healthcare among poor households and to improve the financial sus-tainability of the CBHI scheme (Kalisa et al., 2016). The categorization of households into subsidy groups was based on Ubudehe, a classification system developed by the Rwandan government to categorize all households according to socioeconomic status. This resulted in a stratified premium scheme in which households categorized as having low socioeconomic status received fully subsidized premiums whereas relatively wealth-ier households were subject to a price increase. Exploiting the variation in premium costs created by the policy reform, I estimate the price elasticity of insurance demand using a linear probability model with individual fixed effects. Knowledge of the price sensitivity of demand can inform policy makers regarding the effciency of premium subsidies as a policy tool to promote universal health insurance.

I use panel data from the Rwandan Integrated Household Living Conditions Survey (EICV) in 2010/11 and 2014 to estimate the price sensitivity of the demand for health insurance. The results indicate that the demand for insurance is sensitive to price change but it is not price elastic.3 An increase of the premium costs by 1,000 Rwandan

franc (RwF, corresponding to approximately USD 1.1), is associated with a decrease in the likelihood of enrolling by 10.9 percentage points (ppt) (15% at the mean). This implies an average elasticity of –0.18, indicating that the change in demand is small in relation to the price change. The estimated elasticity is considerably lower than both the elasticity of the demand for the National Health Insurance Scheme in Ghana esti-mated by Asuming et al. (2017) and the price elasticity of demand for preventive health products such as bed nets and deworming medicine (Dupas, 2011; Kremer & Miguel, 2007; Cohen & Dupas, 2010). The effect of changes to premium costs is heterogeneous between different subgroups of households. Individuals living in poor households or households headed by women have a higher price elasticity than individuals in non-poor or male-headed households. As a result of the heterogeneity in price sensitivity, the composition of beneficiaries varies among different subsidy schemes.

The association between household socioeconomic status and insurance premiums, caused by the introduction of the stratified premium scheme, suggests that endogeneity may be a concern for the interpretation of my results. I demonstrate the robustness of my results to omitted variable bias using the Oster (2019) test. Furthermore, I estimate the price sensitivity using samples that are balanced on observable characteristics.

3The demand is considered inelastic if the price elasticity is <|1|; that is, a given percentage

change in the premium cost will cause a smaller percentage change in the demand for insurance.

Comparing individuals with increasingly similar characteristics decreases the concerns that omitted time-varying factors are driving the price sensitivity estimates.

Overall, the results indicate that government subsidy strategies will have a limited effect on insurance coverage. This is a direct effect of the inelastic demand. For example, I find that a decrease in the overall premium costs from RwF 3,000 to RwF 1,000 (USD 3.4 to USD 1.1) would increase take-up from 66% to 77%. Additionally, a subsidy scheme that offers completely subsidized premiums for young children under six years old corresponds to a predicted take-up of 67%. Overall, the simulations indicate that the average insurance coverage remains relatively constant for different subsidy schemes.

Financial sustainability is calculated as the share of insurer cost that is covered by household premiums. I simulate the financial coverage related to the different pricing strategies by calculating the share of insurer costs covered by premium revenue. In ad-dition to considering the potential effects on enrollment levels, this forces me to further consider the association between premium levels and the cost of providing health in-surance. In the presence of selection, changes to the insurance premiums will affect the cost of providing the insurance as the composition of insurance beneficiaries changes in response to the changes in premium costs. As the premium costs increase, so does the cost of providing insurance. Following the analysis presented by Einav et al. (2010), I use unique administrative data on the total costs of providing CBHI in Rwanda and provide evidence of a positive association between insurer costs and premium costs by estimating the average cost curve for administrative sections.4 A positive slope of the

average cost curve indicates that the average insurer cost among enrolled households in a section increases as the average premium level increases, consistent with adverse selection.

I use the association between patient costs and insurance premiums to calculate the financial sustainability in relation to alternative premium schemes. The simulations indicate that the financial coverage of alternative premium subsidies differs depending on whether the insurance market is adversely selected. In the absence of selection, the range of financial coverage levels is wider, between 0.28 and 0.85. In the absence of adverse selection, insurer costs are constant among the different subsidy schemes and variation in the financial coverage is driven by changes in enrollment. Considering the adverse selection scenario, the financial coverage reaches levels between 0.35 and 0.80

4A section is an administrative unit for the CBHI scheme that approximately represents the

(23)

to households with low socioeconomic status. The aim of the targeted subsidy was to increase access to healthcare among poor households and to improve the financial sus-tainability of the CBHI scheme (Kalisa et al., 2016). The categorization of households into subsidy groups was based on Ubudehe, a classification system developed by the Rwandan government to categorize all households according to socioeconomic status. This resulted in a stratified premium scheme in which households categorized as having low socioeconomic status received fully subsidized premiums whereas relatively wealth-ier households were subject to a price increase. Exploiting the variation in premium costs created by the policy reform, I estimate the price elasticity of insurance demand using a linear probability model with individual fixed effects. Knowledge of the price sensitivity of demand can inform policy makers regarding the effciency of premium subsidies as a policy tool to promote universal health insurance.

I use panel data from the Rwandan Integrated Household Living Conditions Survey (EICV) in 2010/11 and 2014 to estimate the price sensitivity of the demand for health insurance. The results indicate that the demand for insurance is sensitive to price change but it is not price elastic.3 An increase of the premium costs by 1,000 Rwandan

franc (RwF, corresponding to approximately USD 1.1), is associated with a decrease in the likelihood of enrolling by 10.9 percentage points (ppt) (15% at the mean). This implies an average elasticity of –0.18, indicating that the change in demand is small in relation to the price change. The estimated elasticity is considerably lower than both the elasticity of the demand for the National Health Insurance Scheme in Ghana esti-mated by Asuming et al. (2017) and the price elasticity of demand for preventive health products such as bed nets and deworming medicine (Dupas, 2011; Kremer & Miguel, 2007; Cohen & Dupas, 2010). The effect of changes to premium costs is heterogeneous between different subgroups of households. Individuals living in poor households or households headed by women have a higher price elasticity than individuals in non-poor or male-headed households. As a result of the heterogeneity in price sensitivity, the composition of beneficiaries varies among different subsidy schemes.

The association between household socioeconomic status and insurance premiums, caused by the introduction of the stratified premium scheme, suggests that endogeneity may be a concern for the interpretation of my results. I demonstrate the robustness of my results to omitted variable bias using the Oster (2019) test. Furthermore, I estimate the price sensitivity using samples that are balanced on observable characteristics.

3The demand is considered inelastic if the price elasticity is <|1|; that is, a given percentage

change in the premium cost will cause a smaller percentage change in the demand for insurance.

Comparing individuals with increasingly similar characteristics decreases the concerns that omitted time-varying factors are driving the price sensitivity estimates.

Overall, the results indicate that government subsidy strategies will have a limited effect on insurance coverage. This is a direct effect of the inelastic demand. For example, I find that a decrease in the overall premium costs from RwF 3,000 to RwF 1,000 (USD 3.4 to USD 1.1) would increase take-up from 66% to 77%. Additionally, a subsidy scheme that offers completely subsidized premiums for young children under six years old corresponds to a predicted take-up of 67%. Overall, the simulations indicate that the average insurance coverage remains relatively constant for different subsidy schemes.

Financial sustainability is calculated as the share of insurer cost that is covered by household premiums. I simulate the financial coverage related to the different pricing strategies by calculating the share of insurer costs covered by premium revenue. In ad-dition to considering the potential effects on enrollment levels, this forces me to further consider the association between premium levels and the cost of providing health in-surance. In the presence of selection, changes to the insurance premiums will affect the cost of providing the insurance as the composition of insurance beneficiaries changes in response to the changes in premium costs. As the premium costs increase, so does the cost of providing insurance. Following the analysis presented by Einav et al. (2010), I use unique administrative data on the total costs of providing CBHI in Rwanda and provide evidence of a positive association between insurer costs and premium costs by estimating the average cost curve for administrative sections.4 A positive slope of the

average cost curve indicates that the average insurer cost among enrolled households in a section increases as the average premium level increases, consistent with adverse selection.

I use the association between patient costs and insurance premiums to calculate the financial sustainability in relation to alternative premium schemes. The simulations indicate that the financial coverage of alternative premium subsidies differs depending on whether the insurance market is adversely selected. In the absence of selection, the range of financial coverage levels is wider, between 0.28 and 0.85. In the absence of adverse selection, insurer costs are constant among the different subsidy schemes and variation in the financial coverage is driven by changes in enrollment. Considering the adverse selection scenario, the financial coverage reaches levels between 0.35 and 0.80

4A section is an administrative unit for the CBHI scheme that approximately represents the

Figur

Updating...

Referenser

Updating...

Relaterade ämnen :